Abstract 67

Pneumonia is a major problem in NICUs and can be caused by the same pathogenic organisms that colonize the lower respiratory tract. This prospective study was undertaken to test the hypotheses that surveillance of airway colonization by BAL would be helpful in predicting a) ventilated infants who developed nosocomial pneumonia and b) pathogens isolated during such episodes. Furthermore, we hypothesized that measurement of interleukin-6(IL-6) in the BAL of colonized neonates would identify those who later developed pneumonia. Eighty intubated neonates underwent BAL 2-3 times a week and, additionally, when pneumonia was inferred. Cultures and IL-6 measurements (ELISA) were done on each BAL. All study infants were receiving antibiotic prophylaxis. A total of 421 BAL procedures were done on the 80 subjects. Forty (50%) infants had growth of microorganisms in one or more BAL cultures. All the BAL cultures yielded one variety of organism only. S. aureus was the most commonly isolated organism. Thirty-four(42%) had late colonization (>72 hours), pneumonia developed in 20 (59%), in 8 (40%) of these was preceded by airway colonization, as demonstrated by BAL positive cultures. There was always agreement between the colonizing organism and those isolated during pneumonia. The sensitivity of routine BAL cultures in predicting neonates who developed ventilator-associated pneumonia was 40%, specificity 76%, PPV 36%, NPV 79%. On the basis of a cutoff level of 200 pg/mL, the sensitivity of the IL-6 measurements in anticipating pneumonia in the colonized infants was 80%, specificity 60%, PPV 66%, NPV 75%. We conclude that surveillance of airway colonization had low sensitivity for predicting nosocomial pneumonia. However, routine BAL cultures were helpful in anticipating pathogens isolated during ventilator-associated pneumonia. IL-6 measurements provided little help for discriminating colonized infants who later developed pneumonia.